Non-normalized Data Research Articles

Validation of a feature-based likelihood ratio method for the SAILR software. Part II: Elemental compositional data for comparison of glass samples

SAILR is open-source software designed to calculate forensic likelihood ratios (LR) from probability distributions of reference data. The purpose of this study was to demonstrate validation of a multivariate feature-based LR method for SAILR using compositional data on glass fragments. Validation was performed using designated performance characteristics, e.g., accuracy, discrimination, and calibration. These characteristics were measured using performance metrics such as cost of the log likelihood ratio and equal error rate. The LR method was developed simultaneously to a baseline method having features less discriminating, but being better aligned with the normality assumption for within-source variation. The baseline method served as the floor of acceptable performance. The results showed that the available data supported LR methods using three elemental features or less. Best performance was obtained using calcium, magnesium, and silicon. The within-source variation in elemental features was slightly leptokurtic (heavy-tailed), violating the assumption of normality. The data were therefore normalized using Lambert W transformation and the performance of the LR method using normalized data was compared with that using non-normalized data. Although performance improved with normalization, the difference was small. Limits of LR output were set to 1/512 ≤ LR ≤ 158 using the empirical lower and upper boundaries (ELUB) LR method. This limited range was primarily a consequence of notable within-source variation. By passing the tests of normality and outperforming the baseline method, the method was considered valid for use in SAILR for data relevant to the background data set, using the defined range of LRs.

Application of a Neural Network Classifier to Radiofrequency-Based Osteopenia/Osteoporosis Screening.

This preliminary study reports application of a neural network classifier to the processing of previously collected data on low power radiofrequency propagation through the wrist with the goal to detect osteoporotic/osteopenic conditions. The data set used includes 67 subjects (23-94 years old, 50 females, 17 males, 27 osteoporotic/osteopenic, 40 healthy). We process the entire spectrum of the propagation coefficient through the wrist from 30 kHz to 2 GHz, with 201 sampling points in total. We found that the dichotomic diagnostic test of raw non-normalized radiofrequency data performed with the trained neural network approaches 90% specificity and ~70% sensitivity. These results are obtained without inclusion of any additional clinical risk factors. They justify that the radio transmission data are usable on their own as a predictor of bone density. With the inclusion of additional clinical risk factors, both specificity and sensitivity improve to 95% and 76% respectively. Our approach correlates well with the available DXA measurements and has the potential for screening patients at risk for fragility fractures, given the ease of implementation and low costs associated with both the technique and the equipment.Clinical Relevance- Dichotomic diagnostic test of raw non-normalized radiofrequency data performed with the trained neural network approaches 90% specificity and ~70% sensitivity. With the inclusion of other clinical risk factors, specificity and sensitivity increase to 95% and 76% respectively.

Including jump height when normalizing single hop impact kinetics can change the directionality of findings

BackgroundAssessing landing kinetics during hop testing could improve return to sport decisions following anterior cruciate ligament reconstruction (ACL) reconstruction. However, different methods for normalizing kinetic outcomes could influence the interpretation of landing kinetics and therefore the clinical recommendations. MethodsTwenty-one females who had returned to sport following primary unilateral ACL reconstructed completed two single hops for maximum distance on each limb. Hop distance, hop height, peak impact force, and impulse were computed for each hop, and peak impact force and impulse magnitudes were assessed when 1) non-normalized 2) normalized by bodyweight, and 3) normalized by peak potential energy during the hop. FindingsAlong with hop distance and height, peak impact force and impulse were found to be lower on the surgical limb relative to the non-surgical limb for both non-normalized data and when normalized to bodyweight only (p < 0.001, d > 0.95). However, peak impact force and impulse were found to be higher on the surgical limb relative to the non-surgical limb when normalizing outcomes to peak potential energy (p < 0.001, d > 1.03). InterpretationDifferent normalization methods result in different interpretations of single hop kinetics. ACL reconstruction patients have shorter hop distances, lower hop heights, lower force magnitudes, and worse energy absorption when hopping on their surgical limb, relative to their non-surgical limb. We believe that normalizing landing kinetics using bodyweight and using peak potential energy provide different information, and as such, we suggest that future research use both methods based on the research question.

Automatic Evaluation of Heart Condition According to the Sounds Emitted and Implementing Six Classification Methods.

The main cause of death in Mexico and the world is heart disease, and it will continue to lead the death rate in the next decade according to data from the World Health Organization (WHO) and the National Institute of Statistics and Geography (INEGI). Therefore, the objective of this work is to implement, compare and evaluate machine learning algorithms that are capable of classifying normal and abnormal heart sounds. Three different sounds were analyzed in this study; normal heart sounds, heart murmur sounds and extra systolic sounds, which were labeled as healthy sounds (normal sounds) and unhealthy sounds (murmur and extra systolic sounds). From these sounds, fifty-two features were calculated to create a numerical dataset; thirty-six statistical features, eight Linear Predictive Coding (LPC) coefficients and eight Cepstral Frequency-Mel Coefficients (MFCC). From this dataset two more were created; one normalized and one standardized. These datasets were analyzed with six classifiers: k-Nearest Neighbors, Naive Bayes, Decision Trees, Logistic Regression, Support Vector Machine and Artificial Neural Networks, all of them were evaluated with six metrics: accuracy, specificity, sensitivity, ROC curve, precision and F1-score, respectively. The performances of all the models were statistically significant, but the models that performed best for this problem were logistic regression for the standardized data set, with a specificity of 0.7500 and a ROC curve of 0.8405, logistic regression for the normalized data set, with a specificity of 0.7083 and a ROC curve of 0.8407, and Support Vector Machine with a lineal kernel for the non-normalized data; with a specificity of 0.6842 and a ROC curve of 0.7703. Both of these metrics are of utmost importance in evaluating the performance of computer-assisted diagnostic systems.

Residual cyclegan for robust domain transformation of histopathological tissue slides.

Variation between stains in histopathology is commonplace across different medical centers. This can have a significant effect on the reliability of machine learning algorithms. In this paper, we propose to reduce performance variability by using -consistent generative adversarial (CycleGAN) networks to remove staining variation. We improve upon the regular CycleGAN by incorporating residual learning. We comprehensively evaluate the performance of our stain transformation method and compare its usefulness in addition to extensive data augmentation to enhance the robustness of tissue segmentation algorithms. Our steps are as follows: first, we train a model to perform segmentation on tissue slides from a single source center, while heavily applying augmentations to increase robustness to unseen data. Second, we evaluate and compare the segmentation performance on data from other centers, both with and without applying our CycleGAN stain transformation. We compare segmentation performances in a colon tissue segmentation and kidney tissue segmentation task, covering data from 6 different centers. We show that our transformation method improves the overall Dice coefficient by 9% over the non-normalized target data and by 4% over traditional stain transformation in our colon tissue segmentation task. For kidney segmentation, our residual CycleGAN increases performance by 10% over no transformation and around 2% compared to the non-residual CycleGAN.

Associations of hamstring and triceps surae muscle spasticity and stance phase gait kinematics in children with spastic diplegic cerebral palsy

Clinical decisions on interventions to improve function in children with cerebral palsy (CP) are based, in part, on hypothesized interactions amongst physical signs of CP and functional deficits. However, a knowledge gap exists regarding associations between spasticity and gait function. This study quantified associations of hamstring and triceps surae spasticity with hip, knee and ankle CP gait patterns. This is a cohort study of children and adolescents [n = 51; 31 male; 20 female; spastic diplegia; Gross Motor Function Classification System I (n = 23) and II (n = 28)] who participated in a clinical consult including gait (Motion Analysis, USA) and modified Tardieu scale (MTS) testing (hamstrings, triceps surae). Shape-based clustering was performed on stance phase sagittal hip, knee and ankle patterns using z-normalized and non-normalized data. Linear regression (R, v3.5.0, R Core Team, Austria) was conducted to assess associations between MTS measures and data clusters (α = 0.05). Shape-clustering revealed two hip and three knee and ankle clusters for z-normalized and non-normalized data. Significant associations of hamstring spasticity and joint patterns were observed for z-normalized knee clusters (CKnee A p = 0.002; CKnee B p = 0.006) and interactions amongst non-normalized hip and knee clusters (CHipA:CKnee B p = 0.033). Trends were observed for soleus spasticity and gastrocnemius range of motion angle and non-normalized ankle clusters (CAnkle B p = 0.051; CAnkle B p = 0.053 respectively). Significant associations of early knee extension and hamstring spasticity, observed using shape-clustering of z-normalized data, provide unique information that may inform the identification of individuals most likely to benefit from spasticity management and targets for spasticity management assessment.

Improving Human Activity Monitoring by Imputation of Missing Sensory Data: Experimental Study

The automatic recognition of human activities with sensors available in off-the-shelf mobile devices has been the subject of different research studies in recent years. It may be useful for the monitoring of elderly people to present warning situations, monitoring the activity of sports people, and other possibilities. However, the acquisition of the data from different sensors may fail for different reasons, and the human activities are recognized with better accuracy if the different datasets are fulfilled. This paper focused on two stages of a system for the recognition of human activities: data imputation and data classification. Regarding the data imputation, a methodology for extrapolating the missing samples of a dataset to better recognize the human activities was proposed. The K-Nearest Neighbors (KNN) imputation technique was used to extrapolate the missing samples in dataset captures. Regarding the data classification, the accuracy of the previously implemented method, i.e., Deep Neural Networks (DNN) with normalized and non-normalized data, was improved in relation to the previous results without data imputation.

Abstract 859: Deep learning predicts expression-based molecular subtypes and immune status of urothelial cancer using digital pathology slides

Abstract RNA-sequencing (RNA-seq) has revealed intrinsic subtypes of urothelial carcinoma (UC). We have recently developed a molecular classifier to separate UC tumors to ‘hot' and ‘cold' based on the expression of immune-related genes. Both molecular subtype and immune-status correlate with response to different treatments, however, RNA-seq data is not readily available for all tumors thus limiting our ability to predict treatment response. We hypothesized that a deep learning approach can predict urothelial cancer molecular subtype and classify immune cell levels from pathology tissue slides. To test this hypothesis, we developed a deep learning model to classify urothelial carcinoma subtypes from hematoxylin and eosin (H&amp;E) stained tissue sections using BASE47 luminal-basal classification and our immune-status classifiers based on RNA-seq data from patients in the TCGA dataset as ground truth. Patches of 2048x2048 pixels were generated from 446 whole slide images (WSIs) resulting in a total of 79,053 images. 70% of the patients were used for training, and 15% are being used as a validation set. The remaining 15% of the data will be a blind test set. We used the Resnet-18 architecture, which performed best based on empirical testing (high accuracy while minimizing overfitting). We performed a color normalization step based on the Vahadane algorithm (IEEE Trans Med Imaging. 2016;35(8):1962 71) to analyze the predictive performance and generalizability of the model compared to non-normalized slides. We also used gradient-weighted class activation mapping to highlight the regions of the tissue that may be important for classification. After 20 epochs of training, we achieved a 74% patch prediction accuracy on the validation set with a ROC-AUC of 0.80 for non-normalized immune cell classification. For normalized images, we achieved a 70% patch prediction accuracy on the validation set with a ROC-AUC of 0.75. For the luminal-basal subtype classification task, we achieved a 68% patch wise accuracy on the validation set with a ROC-AUC of 0.76 for normalized data and 68% patch-wise accuracy with a ROC-AUC of 0.74 for non-normalized data. We are currently testing the model on an internal urothelial cancer cohort at Weill Cornell Medicine. Our data demonstrates the utility of deep learning methods in predicting UC molecular subtype and immune status from digital pathology slides. This data has important implications for predicting prognosis and treatment response in UC patients. Citation Format: Matthew Bryan Brendel, Pegah Khosravi, Emily Tse, Lilly Gu, Kyrillus Shohdy, Rohan Bareja, Bhavneet Bhinder, Olivier Elemento, Bishoy Faltas, Iman Hajirasouliha. Deep learning predicts expression-based molecular subtypes and immune status of urothelial cancer using digital pathology slides [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 859.

Recognition of Activities of Daily Living and Environments Using Acoustic Sensors Embedded on Mobile Devices

The identification of Activities of Daily Living (ADL) is intrinsic with the user’s environment recognition. This detection can be executed through standard sensors present in every-day mobile devices. On the one hand, the main proposal is to recognize users’ environment and standing activities. On the other hand, these features are included in a framework for the ADL and environment identification. Therefore, this paper is divided into two parts—firstly, acoustic sensors are used for the collection of data towards the recognition of the environment and, secondly, the information of the environment recognized is fused with the information gathered by motion and magnetic sensors. The environment and ADL recognition are performed by pattern recognition techniques that aim for the development of a system, including data collection, processing, fusion and classification procedures. These classification techniques include distinctive types of Artificial Neural Networks (ANN), analyzing various implementations of ANN and choosing the most suitable for further inclusion in the following different stages of the developed system. The results present 85.89% accuracy using Deep Neural Networks (DNN) with normalized data for the ADL recognition and 86.50% accuracy using Feedforward Neural Networks (FNN) with non-normalized data for environment recognition. Furthermore, the tests conducted present 100% accuracy for standing activities recognition using DNN with normalized data, which is the most suited for the intended purpose.

The advantages of mass normalized electrochemical impedance spectra for the determination of the kinetic parameters of LiMn2O4 cathodes

The electrochemical properties of LiMn2O4 (LMO) based cathode material with conducting polymer binder containing poly-3,4-ethylenedioxythiophene/polystyrene sulfonate (PEDOT:PSS) dispersion and carboxymethylcellulose (СМС) were investigated for electrodes of three different thicknesses. The effect of the thickness of LMO cathode layer on the kinetic parameters of electrode processes (Rct, Dapp, i0) derived from electrochemical impedance spectroscopy (EIS) has been experimentally analyzed. A simple approach for estimating the possible influence of various real surface areas of electrodes with different active masses by plotting the dependencies normalized to the mass of the active component was proposed. This approach assumes that the real active surface area is defined as the total surface of all particles of the active component, and their mass is proportional to the area given that size of active grains is subject to random statistical distribution. It leads to determination of more reliable values of kinetic parameters and gives the base for more reasonable comparison of results of different works, as was illustrated in the example of LMO.It has been shown that the large difference in the values of charge transfer resistance Rct and apparent diffusion coefficients Dapp for electrodes with various thicknesses becomes insignificant after EIS data normalization to mass loading. While the values obtained from as-recorded non-normalized EIS data referred to the visible area of electrode differ considerably, the normalization of EIS spectra to LMO mass and calculation referred to the real surface area estimated from Brunauer-Emmet-Teller (BET) yields closer values.

Non-Parametric Regression Analysis of Diuron and Gabapentin Degradation in Lake Constance Water by Ozonation and Their Toxicity Assessment

Ozone possesses high selectivity in the oxidation of organic pollutants. It actively reacts with electron donating participants which contain π bonds and non-protonated amines groups. The removal efficiency of organic pollutants hugely depends upon the pollutants’ initial concentration and amount of ozone supplied. This study was conducted at Zweckverband Bodensee-Wasserversorgung (Lake Constance Water Supply), Germany. The prime objective of the research was to observe the performance of diuron and gabapentin ozonation for low ozone doses, therefore meeting the real application requirements of the water treatment plant. Thereby, 1 mg·L−1 of the given organic pollutants was chosen for the treatment. The ozone with a dosage of ≈0.68–1.01 mg·L−1 was generated and homogeneously mixed into Lake Constance water in a semi-batch reactor system. The adequate aliquots of diuron/gabapentin were spiked into the homogenous matrix to acquire the desired initial concentration. The effect of ozone dose and reaction time on the degradation of diuron and gabapentin was investigated. Low ozone doses were sufficient for the complete degradation of diuron and gabapentin, although satisfactory total organic carbon (TOC) reduction was not achieved. Nonetheless, the toxicity from ozone treated effluents can be avoided by adjusting treatment conditions. Due to that degradation data obtained did not follow normalization, the non-parametric (non-normalised) data were analysed with a generalised linear regression model for Gaussian and Poisson distribution. Statistical analysis showed that the ozonation treatment of diuron/gabapentin followed the Gaussian model distribution and the degradation data obtained was proven significant using the Kruskal–Wallis test.

Multicenter Validation Study of Quantitative Imaging Mass Spectrometry.

The aim of this study was to assess potential sources of variability in quantitative imaging mass spectrometry (IMS) across multiple sites, analysts, and instruments. A sample from rat liver perfused with clozapine was distributed to three sites for analysis by three analysts using a predefined protocol to standardize the sample preparation, acquisition, and data analysis parameters. In addition, two commonly used approaches to IMS quantification, the mimetic tissue model and dilution series, were used to quantify clozapine and its major metabolite norclozapine in isolated perfused rat liver. The quantification was evaluated in terms of precision and accuracy with comparison to liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). The results of this study showed that, across three analysts with six replicates each, both quantitative IMS methods achieved relative standard deviations in the low teens and accuracies of around 80% compared to LC-MS/MS quantification of adjacent tissue sections. The utility of a homogeneously coated stable-isotopically labeled standard (SIL) for normalization was appraised in terms of its potential to improve precision and accuracy of quantification as well as qualitatively reduce variability in the sample tissue images. SIL normalization had a larger influence on the dilution series, where the use of the internal standard was necessary to achieve accuracy and precision comparable to the non-normalized mimetic tissue model data. Normalization to the internal standard appeared most effective when the intensity ratio of the analyte to internal standard was approximately one, and thus precludes this method as a universal normalization approach for all ions in the acquisition.

User Environment Detection with Acoustic Sensors Embedded on Mobile Devices for the Recognition of Activities of Daily Living

The detection of the environment where the user is located, is of extreme use for the identification of Activities of Daily Living (ADL). ADL can be identified by use of the sensors available in many off-the-shelf mobile devices, including magnetic and motion, and the environment can be also identified using acoustic sensors. The main objective of this study is to recognize the environments and some standing ADL to include in the development of a framework for the recognition of ADL and its environments. The study presented in this paper is divided in two parts: firstly, we discuss the recognition of the environment using acoustic sensors (i.e., microphone), and secondly, we fuse this information with motion and magnetic sensors (i.e., motion and magnetic sensors) for the recognition of standing ADL. The recognition of the environments and the ADL are performed using pattern recognition techniques, in order to develop a system that includes data acquisition,data processing, data fusion, and classification methods. The classification methods explored in this study are composed by different types of Artificial Neural Networks (ANN), comparing the different types of ANN and selecting the best methods to implement in the different stages of the system developed. Conclusions point to the use of Deep Neural Networks (DNN)with normalized data for the identification of ADL with 85.89% of accuracy, the use of Feedforward Neural Networks (FNN) with non-normalized data for the identification of the environments with 86.50% of accuracy, and the use of DNN method with normalized data for the identification of standing activities with 100% of accuracy.

Subcaudate Tractotomy White Matter Anatomy and Variability: A Diffusion Tensor Imaging Study

Background: Subcaudate tractotomy is a functional ablative procedure performed for treatment-resistant psychiatric disease, targeting the white matter tracts ventral to the head of the caudate nucleus. The white matter anatomy of this region has extensive interindividual variability, and this is thought to impact upon the treatment response to ablative and deep brain stimulation procedures. Objective: We aim to assess in detail the white matter tract anatomy and topographic variability underlying subcaudate tractotomy. Methods: 74 non-depressed adult volunteers underwent MRI including diffusion tensor imaging. Individualized regions of interest were formed in both hemispheres using native non-normalized data to simulate a subcaudate tractotomy. Tractography and analysis were performed and the exact tract locations measured including mean distances and standard deviations to assess variability. Results: The cingulum bundle, uncinate fasciculus, corticostriatal fibres, and corticothalamic fibres were consistently demonstrated. The location of the cingulum bundle was consistent across subjects, but there was extensive interindividual variability in the topographic location of the other tracts. Conclusion: We have demonstrated a detailed analysis of the white matter tracts and their anatomical variability underlying subcaudate tractotomy. This has significant implications for neurosurgical targeting.

Evaluation of Electromyogram Time Characteristics of the Wrist Functional Movements for Intuitive Control of Bionic Prosthesis

Background. Time evaluation features of characteristics (TDF-characteristics) of a surface electromyogram when performing the palm functional movements and the possibility of their implementation in the bionic prosthetic limb intuitive control system.Objective. The aim of the paper is to develop the analytical model for evaluating the TDF-characteristics of myographic signals for basic functional movements of the patient’s wrist and fingers, as well as studying the possibilities of its implementation as a basis for non-parametric method of classification.Methods. A one-channel microcontroller based information measuring system was created for the registration of a surface electromyogram. The analytical model for evaluating its TDF-characteristics based on the use of variance and trapezoidal integral features was developed.Results. The developed model was tested using the method of k-NN classification of the measured signal in the analysis of normalized and non-normalized data of recognition of the fingers’ functional movements. The accuracy of movement classification was 86.11 % and is acceptable for use in the development of methods for automatic control in the control systems of bionic prosthetics of the upper limbs or fingers.Conclusions. The analytical model for evaluating the TDF-characteristics of myographic signals for the wrist basic functional movements based on one-channel measuring system and a simple method of machine learning technique were proposed in the study. The proposed model is effective when working with a small set of signal characteristics and a limited amount of input data, and its classification accuracy can be increased by using a wider sample of data for training.

Bivariate functional principal components analysis: considerations for use with multivariate movement signatures in sports biomechanics

Sporting performance is often investigated through graphical observation of key technical variables that are representative of whole movements. The presence of differences between athletes in such variables has led to terms such as movement signatures being used. These signatures can be multivariate (multiple time-series observed concurrently), and also be composed of variables measured relative to different scales. Analytical techniques from areas of statistics such as Functional Data Analysis (FDA) present a practical alternative for analysing multivariate signatures. When applied to concurrent bivariate time-series multivariate functional principal components analysis (referred to as bivariate fPCA or bfPCA in this paper) has demonstrated preliminary application in biomechanical contexts. Despite this, given the infancy of bfPCA in sports biomechanics there are still necessary considerations for its use with non-conventional or complex bivariate structures. This paper focuses on the application of bfPCA to the force-angle graph in on-water rowing, which is a bivariate structure composed of variables with different units. A normalisation approach is proposed to investigate and standardise differences in variability between the two variables. The results of bfPCA applied to the non-normalised data and normalised data are then compared. Considerations and recommendations for the application of bfPCA in this context are also provided.

Relationship Between Body Fat And Selected Gait Parameters

Over the years obesity has become a widespread pandemic that has affected many people. Obese persons experience difficulty ambulating and conducting daily activities. While body mass indexing is typically used to identify obese individuals percent body fat was preferred for this study because body mass index presents as an inaccurate obesity classification (Shah, 2012). It has been proposed that as percent body fat increases joint mechanics of the hip,knee, and ankle will increase as well on all three planes. PURPOSE: The purpose of this study is to show the correlation between body percent fat and lower extremity joint mechanics. METHODS: 54 participants (24 male, 30 female; Height:170.9±9.2 cm, Weight: 74.4±18.8 kg) were fit with reflective markers on the bony landmarks of the lower extremity and asked to walk across Bertec force plates (Bertec, Inc, Columbus, OH) a total of 10 times. Kinematic data were collected and normalized to body weight and collected with Vicon Nexus 2.4 (Vicon, Inc., Oxford, UK) and processed through Visual 3D software (C-Motion, Inc., Germantown, MD). Body composition was measured using a Bod Pod (CosmedUSE, Inc., Chicago, IL). Data were analyzed using a Pearson correlation coefficients (α=0.05). RESULTS: The data revealed a significant relationship between percent body fat and stride width (r2 = 0.368) as well as peak ankle plantarflexion moment (r2=0.373) (p=0.032 and p=0.005 respectively). No other significant relationships were observed. CONCLUSIONS: Based on this data, step width and peak plantar flexor moments were increased with increased body fat percentage, suggesting that persons with a larger percent body fat require more force to propel themselves forward to toe off. Strengthening plantar flexors, or reducing body fat percentage, may allow those with high percent body fat to ambulate and perform other daily tasks efficiently. Further research is needed to examine non-normalized data to show practical effects. Additionally, the current study used primarily active individuals, a comparison to sedentary individuals is needed.

Investigation of hydrodynamics of gas-solid fluidized beds using cross recurrence quantification analysis

Hydrodynamics of gas-solid fluidized bed was investigated by analyzing its pressure fluctuations using cross recurrence quantification analysis (CRQA). Pressure fluctuations were measured in a lab scale fluidized bed of various particle sizes at different gas velocities. First, the CRQA was applied to a number of well-known dynamic systems and the results demonstrated that it is a powerful method to detect similarities between nonlinear signals. Then, it was shown that graphical structures within the cross recurrence plot of pressure fluctuations of a fluidized bed vary with both superficial gas velocity and particle size. It was found that determinism and cross recurrence rate of non-normalized data initially decrease and then increase with increasing the gas velocity. When the signal is initially normalized, determinism and entropy do not change with the superficial gas velocity while cross recurrence rate is sensible to changes in the superficial gas velocity. It was concluded that entropy can be used for detecting changes of particle size and if a proper reference state is chosen, entropy can be a powerful index for detecting changes in the size of particles in a fluidized bed.

Online Detection of Hydrodynamic Changes in Fluidized Bed using Cross Average Diagonal Line

Online detection of hydrodynamics of gas-solid fluidized bed was characterized using pressure fluctuations by cross recurrence plot (CRP) and cross recurrence quantification analysis (CRQA). Experiments were conducted in a lab scale fluidized bed of various particle sizes 150 μm, 280 μm and 490 μm at different gas velocities. Firstly, pattern changes of cross recurrence plot were discussed and then reference states was selected. Afterwards, cross average diagonal line (CLave) of other states corresponding to reference states were obtained. It was found that cross average diagonal line of non-normalized data initially decreases and then increases with increasing the gas velocity. When the signal is initially normalized, cross average diagonal line does not change with the superficial gas velocity. It was concluded that cross average diagonal line could be used for detecting small changes of particle size and if a proper reference state is chosen, it can be perceived as a powerful index for detecting changes in the size of particles in a fluidized bed.

Electromyographic normalization of vastus lateralis and biceps femoris co-contraction during gait of elderly females

Abstract Introduction: Analyze muscle co-contraction using electromyographic signals, which are normalized to compare individuals, muscles and studies. Maximum voluntary isometric contraction (MVIC) and peak electrical activity (PEA) during movement are the most widely used forms of normalization. Objective: Compare inter-subject variability and investigate the association between the co-contraction indices of the vastus lateralis and biceps femoris during gait, normalized by MVIC and PEA. Methods: Thirty elderly women, aged 70.33 ± 3.69 years took part. Electrical muscle activity during MVIC and gait was recorded using a Biopac MP100 electromyograph. MVIC was performed in a Biodex isokinetic dynamometer. For normalization, the signals were divided by the Root Mean Square values of MVIC and PEA of gait. Results: The coefficient of variation of non-normalized data was 69.3%, and those normalized by PEA and MVIC were 30.4% and 48.9% respectively. Linear regression analysis resulted in a prediction model: PEA = 0.04 + 0.16 x MVIC. The goodness of fit of the regression model was statistically significant (p=0.02). The confidence interval (95% CI) for the intercept was between 0.02 and 0.29 and for MVIC between 0.03 and 0.06. Conclusions: The data normalized by PEA showed less variation than those normalized by MVIC. A 100% variation in data normalized by MVIC resulted in a 16% variation in data normalized by PEA, while variation in normalization by MVIC accounts for 17% of the variation in normalization by PEA and vice versa.